Class BaseNeuralNetwork

Deep Learning with Tensorflow

implements: {TensorScriptModelInterface}

Hierarchy

TensorScriptModelInterface
- BaseNeuralNetwork

Index

Constructors

constructor

Properties

Methods

Constructors

constructor

new BaseNeuralNetwork(options?: TensorScriptOptions, properties?: TensorScriptProperties): BaseNeuralNetwork

Overrides TensorScriptModelInterface.constructor
- Defined in src/base_neural_network.ts:8
Parameters
- Default value options: TensorScriptOptions = {}
  
  neural network configuration and tensorflow model hyperparameters
- Optional properties: TensorScriptProperties
  
  extra instance properties
Returns BaseNeuralNetwork

Properties

compiled

compiled: boolean

getInputShape

getInputShape: (...args: any[]) => any

Type declaration

- (...args: any[]): any
- Parameters
  - Rest ...args: any[]
  Returns any

Optional getTimeseriesShape

getTimeseriesShape: undefined | ((x_timeseries: NestedArray<any> | undefined) => Shape)

Optional layers

layers: TensorScriptLayers | TensorScriptSavedLayers

Optional loss

loss: undefined | number

model

model: any

reshape

reshape: (...args: any[]) => any

Type declaration

- (...args: any[]): any
- Parameters
  - Rest ...args: any[]
  Returns any

settings

settings: TensorScriptOptions

tf

tf: any

tokenizer

tokenizer: any

trained

trained: boolean

type

type: string

Optional xShape

xShape: number[]

Optional yShape

yShape: number[]

Methods

calculate

calculate(input_matrix: Matrix | Vector, options?: PredictionOptions): any

Overrides TensorScriptModelInterface.calculate
- Defined in src/base_neural_network.ts:77
Predicts new dependent variables

override

Parameters
- input_matrix: Matrix | Vector
- Optional options: PredictionOptions
  
  model prediction options
Returns any

returns tensorflow prediction

exportConfiguration

exportConfiguration(): TensorScriptContext

Inherited from TensorScriptModelInterface.exportConfiguration
- Defined in src/model_interface.ts:352
Returns TensorScriptContext

generateLayers

generateLayers(x_matrix: Matrix, y_matrix: Matrix, layers?: TensorScriptLayers | TensorScriptSavedLayers, x_test?: Matrix, y_test?: Matrix): void

- Defined in src/base_neural_network.ts:37
Adds dense layers to tensorflow model

abstract

Parameters
- x_matrix: Matrix
  
  independent variables
- y_matrix: Matrix
  
  dependent variables
- Optional layers: TensorScriptLayers | TensorScriptSavedLayers
  
  model dense layer parameters
- Optional x_test: Matrix
- Optional y_test: Matrix
Returns void

importConfiguration

importConfiguration(configuration: TensorScriptContext): void

Inherited from TensorScriptModelInterface.importConfiguration
- Defined in src/model_interface.ts:363
Parameters
- configuration: TensorScriptContext
Returns void

loadModel

loadModel(options: string): Promise<any>

Inherited from TensorScriptModelInterface.loadModel
- Defined in src/model_interface.ts:402
Loads a saved tensoflow / keras model, this is an alias for

see

https://www.tensorflow.org/js/guide/save_load#loading_a_tfmodel

Parameters
- options: string
  
  tensorflow load model options
Returns Promise<any>

tensorflow model

predict

predict(input_matrix?: Matrix | Vector | InputTextArray | PredictionOptions, options?: PredictionOptions): Promise<any>

Inherited from TensorScriptModelInterface.predict
- Defined in src/model_interface.ts:429
Returns prediction values from tensorflow model

Parameters
- Optional input_matrix: Matrix | Vector | InputTextArray | PredictionOptions
  
  new test independent variables
- Optional options: PredictionOptions
Returns Promise<any>

predicted model values

saveModel

saveModel(options: string): Promise<any>

Inherited from TensorScriptModelInterface.saveModel
- Defined in src/model_interface.ts:415
saves a tensorflow model, this is an alias for

see

https://www.tensorflow.org/js/guide/save_load#save_a_tfmodel

Parameters
- options: string
  
  tensorflow save model options
Returns Promise<any>

tensorflow model

train

train(x_matrix: Matrix, y_matrix: Matrix, layers?: TensorScriptLayers, x_test?: Matrix, y_test?: Matrix): Promise<any>

Overrides TensorScriptModelInterface.train
- Defined in src/base_neural_network.ts:51
Asynchronously trains tensorflow model

override

Parameters
- x_matrix: Matrix
  
  independent variables
- y_matrix: Matrix
  
  dependent variables
- Optional layers: TensorScriptLayers
  
  array of model dense layer parameters
- Optional x_test: Matrix
- Optional y_test: Matrix
Returns Promise<any>

returns trained tensorflow model

Static getInputShape

getInputShape(matrix?: any): Shape

Inherited from TensorScriptModelInterface.getInputShape
- Defined in src/model_interface.ts:333
Returns the shape of an input matrix

function

example

const input = [ [ 0, 1, ], [ 1, 0, ], ]; TensorScriptModelInterface.getInputShape(input) // => [2,2]

see

{https://stackoverflow.com/questions/10237615/get-size-of-dimensions-in-array}

Parameters
- Default value matrix: any = []
  
  input matrix
Returns Shape

returns the shape of a matrix (e.g. [2,2])

Static reshape

reshape(array: Vector, shape: Shape): Matrix | Vector

Inherited from TensorScriptModelInterface.reshape
- Defined in src/model_interface.ts:268
Reshapes an array

function

example

const array = [ 0, 1, 1, 0, ]; const shape = [2,2]; TensorScriptModelInterface.reshape(array,shape) // => [ [ 0, 1, ], [ 1, 0, ], ];

Parameters
- array: Vector
  
  input array
- shape: Shape
  
  shape array
Returns Matrix | Vector

returns a matrix with the defined shape